Solar magnetic tracking. I. Software comparison and recommended practices

Feature tracking and recognition are increasingly common tools for data analysis, but are typically implemented on an ad hoc basis by individual research groups, limiting the usefulness of derived results when selection effects and algorithmic differences are not controlled. Specific results that are affected include the solar magnetic turnover time, the distributions of sizes, strengths, and lifetimes of magnetic features, and the physics of both small scale flux emergence and the small-scale dynamo. In this paper, we present the results of a detailed comparison between four tracking codes applied to a single set of data from SOHO/MDI, describe the interplay between desired tracking behavior and parameterization tracking algorithms, and make recommendations for feature selection and tracking practice in future work.